Simulation of Flow and Heat Transfer of Humid Air in … heat transfer are needed for more power ......

Abstract— This paper reports a CFD investigation on flow

and heat transfer of the humid air which circulates above water

ponds in large installations such as spent nuclear fuel cooling

ponds. The numerical methodology involves a 3-cell zone

approach to enable evaporative boundary condition to be

implemented. Calculations are carried out for seasonal

variations of temperature and humidity which show the

contributions of sensible and latent heat losses. The results

presented in this paper provide useful insights into the flow

development and heat transfer mechanisms present. The results

may be of use in the design of ventilation systems and the

description of the methodology may be useful to practitioners

wishing to carryout similar investigations.

Index Terms—Cooling pond, heat loss, humid air, CFD

I. INTRODUCTION

n many water pool applications such as swimming pools

and Spent Nuclear Fuel (SNF) ponds [1], evaporation from

the free water surface plays a vital role in establishing the heat

transfer and flow field. Detailed data for overall flow field

and heat transfer are needed for more power efficient

swimming pool design [2], and obtaining a longer and safer

operation condition for SNF ponds [3]. Hence it is very

important to understand the mechanism of humid air

distribution to be able to evaluate the amount of heat and

mass loss from such ponds.

Establishing a valid numerical model using CFD is the

main objective of this work in order to simulate the vapour

movement and to estimate the heat transfer in terms of

sensible and latent heat. Furthermore, a parametric study on

the effect of the seasonal climate on flow field and heat

transfer process will be performed based on the typical UK

average weather conditions. The introduced methodology

indicates that CFD can be a reliable tool to offer a very

realistic simulation which can lead to deeper understanding

of complex water vapour behaviour. Also, the method can

be used to evaluate the amount of heat transfer to provide

data for assisting in ventilation system design and the risk

associated with the excessive heating for SNF ponds.

The model is an extension of our previous work presented

earlier [1] and consists of two water pools connected to each

Manuscript received March 17 2014; revised April 1 2014

R. Hasan is with the University of Northumbria, Newcastle, NE1 8ST, UK (corresponding author to provide phone: 0191 243 7233; fax: 0191 227

3684; (e-mail: [email protected]).

J. Tudor is with the University of Northumbria, Newcastle, NE1 8ST, UK (e-mail: [email protected]).

A. Ramadan was a student with the University of Northumbria,

Newcastle, NE1 8ST.

other by a narrow bypass. The main difference in the present

investigation is that we have now focused our attention on

the volume of air on top of the water surface and have

included a ventilation inlet and outlet Fig. 1. Also, the

geometry is a scaled down version of the SNF pool to aid us

in establishing the methodology. These ponds can be

described as a large water basin with concrete walls and

floor, where the evaporation takes place from the free heated

water surface. These two ponds are covered by a concrete

building. In this paper, we propose a 3-cell zone approach to

implement evaporative boundary condition. After a careful

validation exercise we have performed parametric studies on

flow and heat transfer for various seasonal data. Although

the calculations were carried out in the context of SNF, the

methodology is equally applicable for similar other

applications such as leisure swimming pools.

II. METHODOLOGY

A. Steady State Simulation

Steady state simulation was accomplished using the

commercial CFD package of ANSYS FLUENT 14.5 [4].

The general methodology is well established and can be

found in many textbooks such as Versteeg et al. [5]. Navier-

Stokes equations were solved using the SIMPLE algorithm.

In addition, energy equations were solved to take into

account the heat transfer and buoyancy forces were enabled

through momentum equations. Since the air considered is

humid, species transport (scalar) equation has also been

solved. In order to add the turbulence effect, the k-ε model

with standard wall functions has been used.

The boundary condition at the ventilation inlet was

defined as a uniform velocity with magnitude of 0.8 m/s in

X-direction (representing 0.224 Air Change per Hour,

ACH), and temperature equal to the ambient value. The

outflow boundary condition was imposed at ventilation exit.

For wall surfaces, the convection boundary condition was

applied with heat transfer coefficient of 0.25 W/m2-K. The

side and bottom walls of the vapour source zone were

considered to be at constant temperature of 26 oC which was

the same as the water temperature. For all the solid zone

surfaces the no-slip boundary condition was employed.

Non-uniform hexahedral mesh was generated to discretise

the domain and a typical grid distribution is shown in Fig. 2.

The grids were concentrated in areas of steep gradients and,

after a few trials, approximately 0.5 million cells were found

Simulation of Flow and Heat Transfer of Humid

Air in Spent Fuel Cooling Ponds

Ahmed Ramadan, Reaz Hasan* and Jenna Tudor

I

Proceedings of the World Congress on Engineering 2014 Vol II, WCE 2014, July 2 - 4, 2014, London, U.K.

ISBN: 978-988-19253-5-0 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)

WCE 2014

Fig. 1. The 3-D sketch for computational domain.

to yield satisfactory results (that were fairly insensitive to

further resolution) and hence has been used for all the

calculations presented in this paper. For convection terms,

second order upwind differencing has been followed. A

typical computational time on a single Intel core 2Duo

E6600 2.4 GHz processor took about 4 hours in order to

achieve the convergence residual of 10-4

.

B. Implementation of multi-zone flow model

A multi-zone flow model has been incorporated into the

simulation which is explained by reference to Fig. 3 below.

We have considered the domain to consist of three cell

zones: solid zone for the walls, vapour zone for the water

vapour source and humid air zone. Inclusion of these three

zones allows us to implement the boundary conditions in a

very realistic way and also enables us to include variation of

ambient conditions for different seasons.

For the solid zone, the walls were considered to be made of

concrete with thermal conductivity of 1.4 W/m-K. The

vapour zone is a very thin zone (typically three layers of

Fig. 2. Section view of grid distribution for computational domain.

cells) which represents the boundary layer just above the

water surface. Air is considered to be fully saturated with

water vapour at that temperature. This thin boundary layer

was introduced to the continuity equation (1) as mass source

term Sm with water vapour mass flow rate of m = mo – mi ,

where mi represents the absolute humidity at entry

(determined from ambient condition) and mo represents the

absolute humidity at exit (assumed fully saturated).

mSvt

(1)

To maintain dimensional homogeneity, the mass source

term is obtained by (2), where V represents the cell zone

volume.

V

mSm

(2)

III. VALIDATION OF THE METHODOLOGY

The methodology described above was validated against

the published work of Li et al. [6]. The flow domain

involved the movement of humid air in a swimming pool of

practical size (22.5m x 9m x 7.5m). Most simulation

boundary condition and geometry details were available and

implemented in our computations. Figs. 4 and 5 show the

comparisons of temperature and relative humidity (RH)

distributions on the same scale (both shown in grayscale).

Reasonable agreements can be seen for both plots and the

differences may be attributed to slight uncertainty in

geometry specification [6] and turbulence models.

IV. RESULTS AND DISCUSSION

Calculations were carried out for four different seasonal

variations and the results were analysed by plotting velocity,

Fig. 3. Schematic of the three domains.



WCE 2014

temperature, relative humidity and integral values of latent

and sensible heats. Typical flow fields for winter conditions

(temperature: 5oC and RH:85%) are presented and discussed

in this section.

A. Velocity

Fig. 6. Shows the streamlines of humid air on the Z-mid

plane and shows a number of vortical structures. The first

vortex is due to the interaction between the incoming

ventilation air with buoyancy driven flow due to heat

released from the water surface. On the other hand, the

central counter-rotating vortices are a direct manifestation of

the current geometry where there is an ‘inactive’ separation

region between the two ponds. The other vortex near the exit

is just opposite to the inlet vortex. The velocity contours in

Fig. 7 clearly show that magnitudes are very small and most

of the higher velocities are concentrated near the

boundaries.

Fig. 7. Velocity contours on Z-0.4 and Z-2.3 planes.

B. Temperature

Temperature profiles at a few chosen locations in the X-

direction and along the Z-mid plane are shown in Fig. 8. It

can be observed that the temperature is slightly higher above

the water surface than at other places. The effect of the

geometry is visible on the X-mid plane.

Fig. 5(a). RH contour for the present method

Fig. 5(b). R.H contour for Li et al. [6]

Fig. 6. Streamlines on z- mid plane (aspect ratio modified for clarity).

Fig. 8. Temp. contours on Z-mid, X-4, X-mid and X-12 planes.

Fig. 4(b). Temperature distribution for Li et al. [6]

Fig. 4(a). Temperature distribution for the present method.



WCE 2014

C. Relative humidity

Figs. 9 and 10 illustrate the relative humidity distribution

in the X-direction as well as on the Z-mid plane. The rise of

water vapour from the saturated vapour zone is clearly

visible and as it moves upward, a plume of constant high

RH zone is observed. Interestingly, the lower values of RH

display a ‘mushroom’ shape due to interaction with the

ceiling. The RH plots in Fig. 10 shows the distributions on

three Y-planes and are in line with the previous plots.

Fig. 9. R.H. contours on Z-mid, X-4, X-mid, and X-12.

Fig. 10. R.H contours on Y-0.2, Y-1 and Y-1.8.

D. Parametric study

Heat losses through the building were split as sensible

and latent heats and were calculated for different seasonal

variations. The climatic data were incorporated through

ambient air temperature and relative humidity as shown in

Table 1.

E. Seasonal comparison of Heat loss

A comparison of heat loss was conducted for four seasons

and Fig. 11 illustrates the amount of heat loss from the

building body in the form of sensible and latent heat. It can

be clearly seen that the outdoor temperature variation has a

significant effect on the total heat loss through sensible heat.

On the other hand, the moisture content has relatively lesser

effect but still is significant and varies from 18% to 33% of

the total heat loss. We have carried out further parametric

studies by varying the inlet velocity (i.e. ACH) and it was

found that the loss due to latent heat is almost directly

proportional to the ACH while sensible heat loss is rather

insensitive.

Since the evaporation rate is highly dependent on ambient

air temperature and relative humidity [7, 8], a scenario as in

summer configuration, where the ability of the building to

discharge heat is reduced, might cause a drop in the

evaporation rate. Hence the cooling performance in such

ponds would be reduced and would require less energy.

However, for SNF cooling ponds, this observation may be a

cause for concern because it may lead to an undesirable

increase of temperature of the water in which the spent fuels

are kept immersed.

F. Thermal stratification

Fig. 12 shows the vertical temperature gradient on the Z-

mid plane at different X-positions along the length of the

pool. Except positions X-1 and X-15 which are affected by

proximity to inlet and outlet respectively, the curves display

a very slight drop in bulk temperature. Based on the volume

averaged temperature very near to the ceiling and that on the

water surface, we have noted slight temperature

stratification and the values are shown in Table 1. Also

shown on this Table are the numerical values of sensible and

latent heat quantities. Further parametric studies were carried out by varying the

ventilation inlet and outlet conditions in terms of their

location and ACH. As expected, very significant variations

were observed and our findings were mostly in line with

those reported by Said et al. [9] which indicates that opening

the building doors is the main cause in change of the

temperature. In the present case, ventilation inlet and outlet

can be assumed to have the similar effect. For instance, Fig.

13 illustrates the affect in the temperature distribution by

increasing the ventilation inlet velocity to 5 m/s.

Fig. 11. Sensible and latent heat loss variation under different ambient

conditions.

TABLE I Effect of seasonal atmospheric condition on heat transfer and temperature

stratification.

Season Temp.

(oC)

RH

(%)

Strat. Level

(oC)

Sensible

heat loss (W)

Latent heat

loss (W)

Winter 5 85 0.65 625.3 313

Spring 15 60 0.36 326.2 73 Summer 20 76 0.2 177.5 75 Autumn 10 87 0.5 475.4 232



WCE 2014

V. CONCLUSION AND FUTURE WORK

1) Employing the three cell zones in the model allows for

implementing the boundary conditions in a very

realistic and generalised way.

2) The model can be used to predict both sensible and

latent heat for different seasonal conditions.

3) Temperature stratification in the vertical direction has

been found to be small but not insignificant.

4) The boundary conditions need further improvement.

This can be done by considering condensation on the

walls.

5) In order to get more realistic simulation for specific

applications, accurate data about operational and

environmental conditions needs to be collected.

6) Having established the methodology, work is currently

underway to simulate the flow and heat transfer on a

full scale model of SNF pond.

REFERENCES

[1] Robinson, P., R. Hasan, and J. Tudor. Numerical Simulation of Flow and Heat Transfer in Spent Fuel Cooling Ponds. in Lecture Notes in

Engineering and Computer Science: Proceedings of the World

Congress on Engineering. 2013, pp. 1861-1864. [2] Shah, M.M., Prediction of evaporation from occupied indoor

swimming pools. Energy and buildings, 2003. 35(7): p. 707-713.

[3] Kozlov, Y.V., et al., Long-term storage and shipment of spent nuclear

fuel. Atomic Energy, 2000. 89(4): p. 792-803. [4] ANSYS FLUENT 14.5 (2014).

[5] Versteeg, H.K. and W. Malalasekera, An introduction to computational

fluid dynamics: the finite volume method. 2007: Pearson Education. [6] Li, Z., CFD Simulations for Water Evaporation and Airflow Movement

in Swimming Baths. VENTInet, 2005(14): p. 10-11.

[7] Vinnichenko, N.A., et al., Direct computation of evaporation rate at the surface of swimming pool. Recent Researches in Mechanics, 2O11,

2011: p. 120-124.

[8] M. D. Hancock, “Indoor swimming pools and leisure centres – A model to improve operational effectiveness and reduce environmental

impact,” CIBSE Technical Symposium, Sept. 6-7, 2011, De Montfort

University, Leicester, UK. [9] Said, M., R. MacDonald, and G. Durrant, Measurement of thermal

stratification in large single-cell buildings. Energy and buildings,

1996. 24(2): p. 105-115.

Fig. 12. Vertical temperature gradient on Z-mid plane.

Fig. 13. Temperature gradient on Z-mid and X-mid planes for inlet of 5 m/s.



WCE 2014

Simulation of Flow and Heat Transfer of Humid Air in … heat transfer are needed for more power ......

Documents

Transcript of Simulation of Flow and Heat Transfer of Humid Air in … heat transfer are needed for more power ......